[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 Dominique d'Humieres dominiq at lps dot ens.fr changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #19 from Dominique d'Humieres dominiq at lps dot ens.fr --- AFAICT this is fixed by revision 205959: http://gcc.gnu.org/viewcvs/gcc?view=revisionsortby=daterevision=205959 Author:amker Date:Fri Dec 13 11:36:22 2013 UTC (28 hours, 17 minutes ago) Changed paths:8 Log Message: PR tree-optimization/58296 PR tree-optimization/41488 * tree-scalar-evolution.c: Include necessary header files. (simplify_peeled_chrec): New function. (analyze_evolution_in_loop): New static variable. Call simplify_peeled_chrec. * tree-ssa-loop-ivopts.c (mark_bivs): Don't mark peeled IV as biv. (add_old_iv_candidates): Don't add candidate for peeled IV. * tree-affine.h (aff_combination_zero_p): New function. PR tree-optimization/58296 PR tree-optimization/41488 * gcc.dg/tree-ssa/scev-7.c: New test. * gcc.dg/pr41488.c: New test. * g++.dg/pr59445.C: New test. Closing as FIXED.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #17 from Dominique d'Humieres dominiq at lps dot ens.fr --- I fixed the reported problem and posted new patch at http://gcc.gnu.org/ml/gcc-patches/2013-12/msg01159.html Apology that I missed java in bootstrap for previous patch. This version passes bootstrap and test for c,c++,lto,fortran,java,go,objc,obj_c++ on x86_64. Well, it has been a long time without libjava breaking bootstrap! I am not sure if the java case is covered by bootstrap, or other applications. If it's in other application, could anyone help verifying that the issue is addressed on apple-darwin? I have bootstrapped end regtested revision 205924 with the patch on x86_64-apple-darwin13 without any problem.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #18 from bin.cheng amker.cheng at gmail dot com --- Hi Dominique d'Humieres, Thanks for verifying it.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #16 from bin.cheng amker.cheng at gmail dot com --- I fixed the reported problem and posted new patch at http://gcc.gnu.org/ml/gcc-patches/2013-12/msg01159.html Apology that I missed java in bootstrap for previous patch. This version passes bootstrap and test for c,c++,lto,fortran,java,go,objc,obj_c++ on x86_64. I am not sure if the java case is covered by bootstrap, or other applications. If it's in other application, could anyone help verifying that the issue is addressed on apple-darwin? Thanks.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 Richard Biener rguenth at gcc dot gnu.org changed: What|Removed |Added Priority|P3 |P1 Target Milestone|--- |4.9.0
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 Dominique d'Humieres dominiq at lps dot ens.fr changed: What|Removed |Added CC||howarth at nitro dot med.uc.edu --- Comment #1 from Dominique d'Humieres dominiq at lps dot ens.fr --- *** Bug 59455 has been marked as a duplicate of this bug. ***
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 H.J. Lu hjl.tools at gmail dot com changed: What|Removed |Added Status|UNCONFIRMED |WAITING Last reconfirmed||2013-12-10 Ever confirmed|0 |1 --- Comment #2 from H.J. Lu hjl.tools at gmail dot com --- Please add -v to gcj to show what options are passed to jc1
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #3 from Jack Howarth howarth at nitro dot med.uc.edu --- /sw/src/fink.build/gcc49-4.9.0-1000/darwin_objdir/./gcc/jc1 /var/tmp//ccSlyCZfjx -fhash-synchronization -fuse-divide-subroutine -fcheck-references -fuse-boehm-gc -fnon-call-exceptions -fkeep-inline-functions -fPIC -feliminate-unused-debug-symbols -quiet -dumpbase ccSlyCZfjx -mmacosx-version-min=10.8.5 -m32 -mtune=generic -auxbase-strip java/awt/.libs/image.o -g -O2 -Wno-deprecated -version -ffilelist-file -ffloat-store -fomit-frame-pointer -fencoding=UTF-8 -fbootstrap-classes -fsource-filename=/sw/src/fink.build/gcc49-4.9.0-1000/darwin_objdir/x86_64-apple-darwin12.5.0/i386/libjava/classpath/lib/classes -fno-common -fbootclasspath=./:../../../../gcc-4.9-20131210/libjava/classpath/lib/ -faux-classpath ccSlyCZfjx.zip -MD_ -MT java/awt/image.lo -MF java/awt/image.deps -o ccSlyCZfjx.s
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #4 from Jack Howarth howarth at nitro dot med.uc.edu --- (In reply to Jack Howarth from comment #3) /sw/src/fink.build/gcc49-4.9.0-1000/darwin_objdir/./gcc/jc1 /var/tmp//ccSlyCZfjx -fhash-synchronization -fuse-divide-subroutine -fcheck-references -fuse-boehm-gc -fnon-call-exceptions -fkeep-inline-functions -fPIC -feliminate-unused-debug-symbols -quiet -dumpbase ccSlyCZfjx -mmacosx-version-min=10.8.5 -m32 -mtune=generic -auxbase-strip java/awt/.libs/image.o -g -O2 -Wno-deprecated -version -ffilelist-file -ffloat-store -fomit-frame-pointer -fencoding=UTF-8 -fbootstrap-classes -fsource-filename=/sw/src/fink.build/gcc49-4.9.0-1000/darwin_objdir/x86_64- apple-darwin12.5.0/i386/libjava/classpath/lib/classes -fno-common -fbootclasspath=./:../../../../gcc-4.9-20131210/libjava/classpath/lib/ -faux-classpath ccSlyCZfjx.zip -MD_ -MT java/awt/image.lo -MF java/awt/image.deps -o ccSlyCZfjx.s opps, that actually was -mtune=core2 but -mtune=generic also segfaults jc1.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #5 from H.J. Lu hjl.tools at gmail dot com --- (In reply to Jack Howarth from comment #4) (In reply to Jack Howarth from comment #3) /sw/src/fink.build/gcc49-4.9.0-1000/darwin_objdir/./gcc/jc1 /var/tmp//ccSlyCZfjx -fhash-synchronization -fuse-divide-subroutine -fcheck-references -fuse-boehm-gc -fnon-call-exceptions -fkeep-inline-functions -fPIC -feliminate-unused-debug-symbols -quiet -dumpbase ccSlyCZfjx -mmacosx-version-min=10.8.5 -m32 -mtune=generic -auxbase-strip java/awt/.libs/image.o -g -O2 -Wno-deprecated -version -ffilelist-file -ffloat-store -fomit-frame-pointer -fencoding=UTF-8 -fbootstrap-classes -fsource-filename=/sw/src/fink.build/gcc49-4.9.0-1000/darwin_objdir/x86_64- apple-darwin12.5.0/i386/libjava/classpath/lib/classes -fno-common -fbootclasspath=./:../../../../gcc-4.9-20131210/libjava/classpath/lib/ -faux-classpath ccSlyCZfjx.zip -MD_ -MT java/awt/image.lo -MF java/awt/image.deps -o ccSlyCZfjx.s opps, that actually was -mtune=core2 but -mtune=generic also segfaults jc1. What is the equivalent of -march=?
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #6 from Dominique d'Humieres dominiq at lps dot ens.fr --- What is the equivalent of -march=? I configure with --with-arch=corei7 --with-cpu=corei7.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #8 from H.J. Lu hjl.tools at gmail dot com --- (In reply to H.J. Lu from comment #7) classpath/java/awt/image/SinglePixelPackedSampleModel.java is used on Linux. Is there a way to compile it by hand ^ I meant isn't used. on Linux?
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #7 from H.J. Lu hjl.tools at gmail dot com --- classpath/java/awt/image/SinglePixelPackedSampleModel.java is used on Linux. Is there a way to compile it by hand on Linux?
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445
Markus Trippelsdorf octoploid at yandex dot com changed:
What|Removed |Added
CC||octoploid at yandex dot com
--- Comment #9 from Markus Trippelsdorf octoploid at yandex dot com ---
This issue also happens when building LLVM.
Reduced:
markus@x4 llvm_build % cat test.ii
template typename _Iterator struct A;
template typename _Tp struct A_Tp * {
typedef _Tp value_type;
typedef int difference_type;
};
template typename _Compare struct B {};
template typename _Compare struct C {
_Compare _M_comp;
template typename _Value, typename _Iterator
int operator()(_Value p1, _Iterator p2) {
return _M_comp(p1, *p2);
}
};
template typename _Compare C_Compare __val_comp_iter(B_Compare);
template typename _RandomAccessIterator, typename _Compare
void __unguarded_linear_insert(_RandomAccessIterator p1, _Compare p2) {
typename A_RandomAccessIterator::value_type a;
_RandomAccessIterator b = p1;
--b;
while (p2(a, b)) {
*p1 = 0;
p1 = b;
--b;
}
}
template typename _RandomAccessIterator, typename _Compare
void __insertion_sort(_RandomAccessIterator, _Compare p2) {
for (_RandomAccessIterator c;; ++c)
__unguarded_linear_insert(c, __val_comp_iter(p2));
}
template typename _RandomAccessIterator, typename _Distance, typename
_Compare
void __chunk_insertion_sort(_RandomAccessIterator, _Distance, _Compare p3) {
_RandomAccessIterator d;
__insertion_sort(d, p3);
}
template typename _RandomAccessIterator, typename _Pointer, typename _Compare
void __merge_sort_with_buffer(_RandomAccessIterator p1, _Pointer, _Compare p3)
{
__chunk_insertion_sort(p1, 0, p3);
}
template typename _RandomAccessIterator, typename _Pointer, typename
_Distance,
typename _Compare
void __stable_sort_adaptive(_RandomAccessIterator, _Pointer, _Distance,
_Compare p4) {
_RandomAccessIterator e;
__merge_sort_with_buffer(e, 0, p4);
}
template typename _RandomAccessIterator, typename _Compare
void __stable_sort(_RandomAccessIterator p1, _Compare p2) {
__stable_sort_adaptive(
p1, 0, typename A_RandomAccessIterator::difference_type(), p2);
}
template typename _RandomAccessIterator, typename _Compare
void stable_sort(_RandomAccessIterator, _RandomAccessIterator p2, _Compare) {
B_Compare f;
__stable_sort(p2, f);
}
class D {
public:
void m_fn1();
};
class F {
struct G {
D MFI;
int operator()(int p1, int p2) {
if (p1)
return 0;
if (p2)
return 1;
MFI.m_fn1();
}
};
void m_fn1(int p1) const;
};
void F::m_fn1(int p1) const {
int *g, *h;
stable_sort(h, g, G());
}
markus@x4 llvm_build % g++ -O2 -c test.ii
test.ii: In member function ‘void F::m_fn1(int) const’:
test.ii:74:6: internal compiler error: in add_old_iv_candidates, at
tree-ssa-loop-ivopts.c:2541
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 Markus Trippelsdorf octoploid at yandex dot com changed: What|Removed |Added CC||amker.cheng at gmail dot com --- Comment #10 from Markus Trippelsdorf octoploid at yandex dot com --- Started with r205848.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 H.J. Lu hjl.tools at gmail dot com changed: What|Removed |Added Target|x86_64-apple-darwin13 | Status|WAITING |NEW Host|x86_64-apple-darwin13 | Build|x86_64-apple-darwin13 |
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #11 from Dominique d'Humieres dominiq at lps dot ens.fr --- Started with r205848. Well, the assert at line 2541 has been introduced in the revision: +gcc_assert (gimple_bb (phi) == data-current_loop-header); and it was in the original patch et http://gcc.gnu.org/ml/gcc-patches/2013-12/msg00630.html, but without explanation for it.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #12 from Jack Howarth howarth at nitro dot med.uc.edu --- Added Jeff Law since he reviewed and approved the offending patch.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #13 from bin.cheng amker.cheng at gmail dot com --- Sorry for bothering, I have reverted the patch. Will investigate it.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445
--- Comment #14 from bin.cheng amker.cheng at gmail dot com ---
I found out the root cause of this ICE and will use the simplified code given
by comment#9 as an example.
The gimple dump before IVOPT is like:
bb 2:
bb 3:
# c_2 = PHI c_5(D)(2), c_3(9)
__val_comp_iterF::G (D.4949);
p2 = D.4950;
c_6 = c_2 + 4294967292;
_21 = MEM[(int *)c_2 + 4294967292B];
if (a_11(D) != 0)
goto bb 4;
else
goto bb 5;
bb 4:
c_3 = c_2 + 4;
goto bb 9;
bb 5:
goto bb 7;
bb 6:
# c_23 = PHI c_28(11), c_7(10)
# _24 = PHI _29(11), _22(10)
bb 7:
# c_20 = PHI c_15(6), c_2(5)
# c_15 = PHI c_23(6), c_6(5)
# _26 = PHI _24(6), _21(5)
if (_26 != 0)
goto bb 11;
else
goto bb 8;
bb 8:
D::m_fn1 (MEM[(struct G *)p2].MFI);
if (_13(D) != 0)
goto bb 10;
else
goto bb 4;
bb 9:
goto bb 3;
bb 10:
*c_20 = 0;
c_7 = c_15 + 4294967292;
_22 = *c_7;
goto bb 6;
bb 11:
*c_20 = 0;
c_28 = c_15 + 4294967292;
_29 = *c_28;
goto bb 6;
With the patch:
STEP1: # c_20 = PHI c_15(6), c_2(5) is recognized as an iv.
STEP2: Since # c_15 = PHI c_23(6), c_6(5) comes from a merging conditional
branches, it shouldn't be marked as a biv in mark_bivs.
STEP3: When mark_bivs handling # c_20 = PHI c_15(6), c_2(5),it should know
that this is a peeled iv and not mark either iv(c_20) or incr_iv(c_15) as bivs.
Unfortunately, this patch should add logic in mark_bivs to skip peeled iv,
rather than give an assert later when adding candidates for bivs.
The following patch should fix this problem:
@@ -1074,7 +1074,7 @@ find_bivs (struct ivopts_data *data)
static void
mark_bivs (struct ivopts_data *data)
{
- gimple phi;
+ gimple phi, def;
tree var;
struct iv *iv, *incr_iv;
struct loop *loop = data-current_loop;
@@ -1090,6 +1090,13 @@ mark_bivs (struct ivopts_data *data)
continue;
var = PHI_ARG_DEF_FROM_EDGE (phi, loop_latch_edge (loop));
+ def = SSA_NAME_DEF_STMT (var);
+ /* Don't mark iv peeled from other one as biv. */
+ if (def
+ gimple_code (def) == GIMPLE_PHI
+ gimple_bb (def) == loop-header)
+continue;
+
incr_iv = get_iv (data, var);
if (!incr_iv)
continue;
PS, the example code can be optimized with fixed version patch by recognizing
more address ivs. I attached the generated assembly code for arm cortex-m3.
[Bug tree-optimization/59445] [4.9 Regression] ICE in add_old_iv_candidates, at tree-ssa-loop-ivopts.c:2541
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59445 --- Comment #15 from bin.cheng amker.cheng at gmail dot com --- Created attachment 31414 -- http://gcc.gnu.org/bugzilla/attachment.cgi?id=31414action=edit The generated assembly with/without patch for code in comment #9 on cortex-m3
